These days, Artificial Intelligence is the name of the game in the tech world. The rise of AI has contributed to massive growth in the sector, pushing the American economy further at a time when many economists were predicting stagnation or decline. AI serves as a reminder that technology always has the potential to surprise us, and it is one of the few industries capable of radically changing our society in a matter of months or even weeks.

But although AI is now the darling of the tech sector, it is not without its faults. The longer AI remains in the spotlight, the more lawsuits it faces. So far, legal action has failed to dent this new tech juggernaut. But a new lawsuit filed by some of the biggest names in Hollywood throws fresh doubts on the long-term viability of AI in its current form. OpenAI is now being sued for using copyrighted material to train ChatGPT. But what does this mean for technology law?

Sarah Silverman Claims OpenAI Used Her Memoir to Train ChatGPT

Sarah Silverman and other authors have sued both OpenAI and Meta, claiming that their artificial intelligence was trained using copyrighted material. This includes Silverman’s successful memoir, Bedwetter. In one exhibit, Silverman and her attorneys showed that when prompted, OpenAI is capable of summarizing her entire book from start to finish. The argument is that if the software is capable of summarizing the book, it must have access to the entire text. But there is only one problem: OpenAI never asked for permission to use the full text. Silverman and her legal team argue that this is evidence OpenAI accessed an online copy of the book illegally.

Similar examples have been highlighted by two other authors who have joined Silverman’s lawsuit. One is Christopher Golden, who showed that ChatGPT is capable of summarizing his book Ararat in exactly the same way. The third is Richard Kadrey, who showed an identical situation with his book Sandman Slim. These three alleged copyright violations show that there may be many other books that were accessed and “scraped” illegally by OpenAI as they trained ChatGPT.

Meta has also been named as a defendant, as the books were apparently accessed by the tech company as it trained its LLaMA AI models that were recently introduced in February of 2023. The plaintiffs claim that both of these companies used illegal “shadow libraries” to access the books. They specifically mentioned a platform called Bibliotik and claimed that they could trace Meta and OpenAI’s use of this illegal source through its publicly-accessible papers.

As a result, the plaintiffs are claiming various damages – including copyright violations, negligence, unjust enrichment, and unfair competition. As this story is gaining attention from across the world, other authors and publishers are joining the discussion and speaking out about what they believe is a gross violation of their intellectual properties.

In truth, authors only represent one potential group of plaintiffs in this new fight against AI. Artists and programmers are also coming forward, claiming that their copyrighted material was used to train various image-generation tools, chatbots, and programming tools. People often forget that online images are often protected by copyrights. If it is true that AI image generation tools accessed these images without permission, it could lead to major legal consequences. The question is whether or not the plaintiffs can actually prove that this intellectual property violation occurred as the AIs were trained.

Of course, this is nothing new. OpenAI has faced numerous lawsuits despite being in the spotlight for a relatively short time. Meta has seen scores of lawsuits over the years, and the tech company is still standing. That being said, some are wondering whether this particular lawsuit could actually sink the rise of AI. While other AI lawsuits have stood on shaky foundations, this one seems to be based on solid legal issues that are clearly documented. If this lawsuit does cause serious issues for ChatGPT and Meta, it could cause tech companies to completely rethink the way they train, market, and use AIs.

After paying damages to everyone who was affected by this alleged copyright violation, ChatGPT, Meta, and any other company found guilty would need to conduct all future training by carefully getting permission from all creators when training AI using content. The question is whether artists would actually provide this permission. After all, if AI stands to replace countless writers, podcasters, and actors, then isn’t it in their best interest to collectively refuse to provide this permission? Some might cave in and provide access to their content, but only if they receive fair compensation.

We know that AI is already having a serious impact on the entertainment industry. Screenwriters in Canada and the United States recently went on strike, citing fears that AI could eventually take their jobs. Among other things, the Screenwriters’ Guild is demanding that Hollywood immediately cease using AI to generate scripts for some of the most popular TV shows in the world. They also point out that AI is being used to rewrite literary material that scriptwriters have already published – once again echoing copyright concerns. Some argue that it is “just a matter of time” before AI-generated performances replace “flesh and bone” actors entirely.

The concepts are not new but the use of large language data sets culling enormous data sets that may have been authorized with a very different set of processing in mind, undermines the effectiveness of the consent to process that data. Scaping data from the web and using enormous processing power to aggregate, categorize and process that data to derive detailed insights and prospects on people associated with that data, their buying habits and buying preferences is a big and very profitable business and many of the worlds mega processing and hosting companies have the resources and the motivation to challenging the legal roadblocks as the AI market emerges before our eyes.

Faced with these hurdles, there are very real concerns over whether AI can survive as a viable business model. The age of “free content generation” may fade away just as quickly as it appeared. Can the law start to close the gap with this new and rapidly evolving segment of the tech marketplace. Of course, this is all speculation – and the tech industry has shown remarkable resilience against a slew of lawsuits over the years. ChatGPT and other companies exploring the possibilities of AI have every chance of survival – especially if they get help from qualified, experienced technology attorneys.

Where Can I Find an Experienced Technology Lawyer?

 If your startup, SaaS, or tech company is facing legal issues, consider getting in touch with John P. O’Brien. Legal issues involving technology require the attention of lawyers who have experience in this specific sector. Technology often advances faster than the legislation created to legislate it, but this does not make companies immune from lawsuits. To protect your business’s best interests and avoid financial issues, book a consultation today. We will discuss your unique situation and determine the best route forward.